Jump to Navigation

Temperature Measurement

Many sensors can measure temperatures. All sense some change in a physical characteristic. Six common types of temperature sensors are:  thermocouples, resistive temperature devices (RTD’s and thermistors), infrared radiators, bimetallic devices, liquid expansion devices and change-of-state devices.

In the process industries, the commonly used temperature sensors are thermocouples and resistive devices.  Thermocouples are the most often used.  A thermocouple consists of two alloys joined together at one end and open at the other.  A change in temperature at the joined end produces an EMF (electromotive force) across the open end, the output end.  As the temperature goes up, the EMF goes up, although not necessarily in a linear fashion.  The open-end EMF is a function of the closed-end temperature, and the temperature at the open end.  In Figure 1, as the closed end (T1) is heated, the EMF (V1) at T2 will rise.  Only by holding the temperature at T2 can the measured EMF be considered a direct function of the change in T1.  The standard for T2 is 0°C.  Tables and charts used to correlate V1 to T1 assume T2 is at 0°C.  Since this is seldom true, the difference between T2 and 0°C is corrected for in the instrument to which the thermocouple is connected.  This EMF adjustment is called the “cold-junction”.

Figure 1.

Temperature changes in the wiring between the input and output ends do not affect the output voltage, as long as the wiring is of thermocouple alloy or its equivalent.  Of course, this assumes the wire can’t get hot enough to melt or change its electro thermal behavior.

Thermocouple readings are not affected by insertion of non-thermocouple alloys in either one or both leads, so long as the temperature at the ends of the non-thermocouple alloy is the same temperature.  See Figure 2.  As long as both T3’s are at the same temperature, V1 will correspond to T1.  This allows switches to be used in thermocouple transmission wires.  Usually, the wiring is of thermoelectric equivalent of the thermocouple; in other words, a “J” thermocouple uses “J” wire.  If a switch is inserted in the “J” wire, as long as it contacts are gold or silver-plated, springs insure good contact, and the connections are at the same temperature, the switch should not affect the input/output of the thermocouple.

Figure 2.

Figure 3 shows the many types of thermocouples available.  Type J and K are by far the most commonly used.  (There are others, such as M, G, D, etc., but they are seldom-used special non-ANSI thermocouples.)

Figure 3.

Temperature range and the environment are the main considerations when selecting a thermocouple.  Smaller gauges provide faster response time, but shorter service life.  Conversely larger gauges last longer, but give reduced response time.

Figure 4 shows the maximum/minimum temperatures for a specific thermocouple made with different gauges.

Figure 4.

Figure 5 shows color-coding and magnetic properties for thermocouples and thermocouple wire.  This can be useful for identifying in-service thermocouples.

Figure 5.

Thermocouples are usually located inside of the metal or ceramic shields that protect them from a variety of environments. 

Resistive temperature sensing devices utilize the fact that electrical resistance of a material changes as its temperature changes.  Two types are:  resistance temperature detectors, referred to as RTD’s, and thermistors.  RTD’s rely on a resistance change in a metal, the resistance rising fairly linearly with a rise in temperature.  This is known as “positive temperature coefficient” — PTC.

Thermistors are usually a ceramic semiconductor in which the resistance drops nonlinearly with a temperature rise.  This is a “negative temperature coefficient”— NTC.

RTD’s are usually made from a fine platinum wire wrapped around a ceramic or glass core.  In use, a small current is passed across the element and a voltage change is produced, which is proportional to the resistance change caused by varying temperatures.  This change is measured and converted to units of heat calibration.

Most RTD’s are low resistance.  Most common is 100 ohms at 0°C.  The slope of the resistance vs. temperature plot for the RTD is referred to as an alpha value.  Alpha stands for temperature coefficient.  The slope for a given sensor depends on the purity of the platinum.  The most commonly used standard for purity has a value of .00385.  This is known as the “European Curve” or “European Standard”.  There is also an “American Standard”.  The “American Standard” has a higher alpha value of .00392.  If the alpha value for an RTD is not specified, it is assumed to be .00385.

Since RTD’s have an initial low resistance, usually 100 ohms, they exhibit a small change in resistance per unit of temperature change.  See Figure 6.  Because of this, the resistance in the lead wire connecting an RTD to an instrument has to be considered.  While there is two wire RTD’s, the most commonly used is a 3-wire. 

See Figure 7.

Figure 6.

Figure 7.

Three-wire RTD’s provide one connection to one end of the element and two to the other end of the element.  Connected to an instrument designed to accept three-wire input, sufficient compensation is usually achieved for lead wire resistance and temperature change in lead wire resistance.

Four-wire provides two connections to each end of the element to completely compensate for lead wire resistance and temperature change in lead wire resistance.  This configuration is used where highly accurate temperature measurement is vital.

RTD’s are resistive devices that function by passing a current through a sensor.  While this is a very small current, it does create heat and can throw off the temperature reading.  This self-heating may be significant when dealing with a still sensed media.  There is no carry-off of the heat generated.  When used to sense a flowing or agitated media, this heating is negligible in most applications.  (It should be noted here that this problem does not arise with thermocouples because they are almost zero current devices.)

Thermistors share this self-heating phenomenon.  Unlike RTD’s, thermistors have a resistance/temperature relationship that is negative and highly nonlinear.  As the sensed temperature rises, the resistance values decrease.  They are “negative temperature coefficient”— NTC devices.  Thermistors are comparatively inexpensive, but have a limited sensing range of -40°F to +300°F.  Thermistors are usually designed with their resistance reference at 25°C (77°F), and 2252 ohms, 5000 ohms, or 10,000 ohms.  They produce a large resistance change per degree of temperature change.  See Figure 8.  Note the non-linearity of the temperature resistance relationship. While the vast majority of thermistors are NTC, they can be made PTC due to modern technological advances.  Various manufacturers will make special thermistors specific for their systems, even PTC thermistors, such as Johnson Controls 350 system sensors, the SET189A.

Figure 8.

This “silicon” sensor uses 1035 ohms at 77°F for its reference and is almost linear in its temperature response.  Honeywell uses thermistors with varying reference points, such as the C7100A, B, and D.  The “A” at 77°F is 3484 ohms, the “B” 22,800 ohms, and the “D” 1,097 ohms. 

Lead wires for connecting thermistors should be stranded copper wire, shielded or unshielded as required, and wire sized to prevent the added resistance of the wire introducing a reading error. This is usually not a problem in runs of 50 feet or less.

Product Categories:
Feature this resource?: 

Main menu 2